Framework for Detecting Student Behaviour (Nail Biting, Sleep, and Yawn) Using Deep Learning Algorithm

Human action recognition is an intriguing and demanding area of study in computer vision. Robotics, intelligent surveillance, security, healthcare, education, and human-computer interfaces are just a few of the many applications for it. It is possible to evaluate students' attitudes and abilities by seeing and analyzing their behavior in the classroom. Consequently, a great deal of research has used student action recognition to pinpoint student activities. However, most of this study has focused on a specific student behavior, such raising their hands, sleeping, or yawning. For the benefit of students' welfare and education, a framework for recognizing certain behaviors in the classroom must be developed. In this research, we use a specialized kind of deep learning called the CNN algorithm to detect student behavior algorithm was developed mainly for applications requiring the detection, segmentation, and classification of images. CNN are used in security camera systems, and other real-world applications. When a student shows behaviors, such as sleeping, yawning, or nail-biting, the alarm sounds out. It can handle student behavior recognition better and has better detection performance compared with other tiny target CNN detection algorithms and the YOLO technique. Being aware of how students interact with the classroom, recognizing when they are uncomfortable, distracted, or disengaged, and reacting accordingly may end up in an enjoyable learning environment and enhanced educational quality. Using a feature extraction approach in this study, we can create a framework to categorize student behaviors like biting their nails, sleeping, and yawning based on eye and mouth traits.